SUMMARY Hippocampal memory circuits are strongly implicated in the formation and persistence of delusions. In particular, the dentate gyrus (DG) and the CA1 subfield are vulnerable to early developmental stressors that are risk factors for schizophrenia; postmortem evidence points to deficits in inhibitory input and in neurogenesis. We hypothesize that delusions result from aberrant associative memory formation due to impaired functioning of DG/CA3 subfields and persist due to a failure to update false beliefs with new episodic information due to reduced mnemonic prediction error signaling in CA1 and reduced post-encoding consolidation of newly formed memories. Because antipsychotics target hippocampal memory circuits, it is important to study these circuits in unmedicated subjects. We propose to apply three task-based fMRI paradigms to examine early mnemonic associative processing, prediction error, and plasticity of circuits associated with encoding and retrieval in three experiments, each including 50 first episode nonaffective psychosis (FEP) subjects and 50 healthy matched controls. The medication-nave FEP subjects will be re- studied after 8 weeks of antipsychotic treatment to examine the relationship between medication effects on delusional severity and on hippocampal memory circuits. Imaging will also be repeated after 8 weeks in healthy controls to assess learning effects. The first paradigm, a behavioral pattern separation task, has not previously been studied in medication-nave first episode psychosis. The paradigms for assessment of CA1 activation during prediction error and of plasticity of connectivity between hippocampal subfields, the ventral tegmental area dopamine neurons and cortical regions were recently developed and validated by Dr. Davachi's team in healthy subjects; we have demonstrated feasibility of these paradigms in schizophrenia subjects. The proposed project will advance our understanding of circuits involved in delusions and their pharmacologic response, will provide validated imaging biomarkers for clinical studies and will identify new targets for treatment development.